JPS5930151A - System freezing system - Google Patents

Abstract

PURPOSE:To constitute so that any computer can be stopped causing no inconsistency, by executing a unified control by synchronizing with each other, to plural independent computers coupled by an inter-channel coupling device. CONSTITUTION:When a stop request is instructed from a console device 1, a stop processing task 3 transmits stop commands 11, 11' to all other subcomputers 4, 4', and starts stop processing tasks 6, 6'. The stop processing tasks 6, 6' checks a direct access storage device DASD 7 which is used in common between plural computers, and monitors an occupied state of the DASD 7 at a constant time interval. In this state, only when the occupied state of all DASDs is released within a fixed time, the stop processing of a processor is continued, and in case when an unreleased DASD exists, the stop processing is suspended. In this case, a new request to the DASD is all set to a waiting state.

Description

DETAILED DESCRIPTION OF THE INVENTION Object of the Invention The present invention relates to a system freezing method in a complex computer system in which a plurality of independent computers are connected by an inter-channel coupling device.

Prior art The conventional system freezing method was effective only in a single 7V computer system, so in the case of a compound computer system, multiple computers share files, so it is difficult to freeze a specific computer system. When frozen, jobs running on the frozen computer system may retain their files and other resources. In this state, if another computer that does not freeze the same file or other resources is used as a resource, the job cannot be executed.

OBJECTS OF THE INVENTION An object of the present invention is to provide a system freezing method for a complex computer system that eliminates drawbacks such as confinement.

According to the system freezing method of the present invention, a plurality of independent computers connected by an inter-channel coupling device synchronize with each other and perform unified fltlJ il, so that no contradiction occurs in any of the 119 computers. Stop at.

That is, according to the present invention, when a system freezing factor occurs,
The main computer, which is responsible for monitoring the entire system, issues a stop command to other computers (secondary computer).

The causes of system freezing include a single stop request from an operator or program, detection of a power supply, and detection of a fault condition in other systems (which require a tlj connection). .

When each subordinate 77-Iso receives a stop command 1) from the master d1-1 aircraft, it performs an operation to save the contents of the main memory bjt.
Stop processing! I+f+L, when the process is completed, returns a completion notification to the master J'l, machine. The main controller Ena and Iso monitor each wave meter Jl-C to see if the optical f notification is returned from Iso within the monitoring time, and disconnect the slave computers that do not receive a completion notification within the monitoring time from the system. The master 17-machine then initiates its own shutdown process.

An embodiment of the present invention will be described below with reference to the drawings.

Examples of the invention FIG. 1 is a schematic 'Iq diagram of the belly computer system.

In the figure, 2 is a main computer responsible for monitoring the entire system, and 4.4' is a wave meter W. Although the number of slave computers is arbitrary, only two are shown for convenience.・Main g1
A channel coupling device (C'r CA ) 5.5' connects the degree calculator 2 and each wave meter 4.4'.

7 shows only the main external combinations that are shared by n wave meters, (
The number of doors to be used is arbitrary. 8.10.101
The main storage device (an external storage device for saving the contents of the memory 9) of the main computer 2 and the secondary computer 4.4'.

Immediate access storage (DASIJ) such as magnetic disk drives are commonly used. ■ is a console device.

The processing flow of the processing flow control subcomputer 4.4' of the main IT n Iso 2 related to system freezing is shown in FIGS. 2 and 3, respectively. The system freeze will be explained below with reference to Figures 2 and 3.

System freeze factor, for example, when the operator receives a 01'4G f ff request -fLi through the device I, the stop water retention is the main function. Contact. Stop processing task 3 is
All other wave meter U machines (main calculation Isoi Tato's total 11-Toru) 4
．． 4'(, 'i' via CA 5.5 stop 1
Self<rll, send 1 work, each slave f! l'%, set the monitoring start time for each machine (step 50.51.52
),, and notification of the completion of the stop processing of Seisho 1 Stone Excerpt 5.5 or E is performed at 12.120 o'clock music monitoring (Step 5)
4.55).

1l)) - Machine 4 (same as 4')
The 1 stop command 11 (Il'j) sent via the CA5 (5') is sent to the stop processing task 6 (6') of the wave meter 1'1 machine, the slave computer 4 (4'). ) stop processing χ starts 1°.

Stop Jl! +! Task 6 is to check S 1) 7 to D that is shared among multiple machines.
RIG 5ERVE status (exclusive status of IJAsIJ by hardware interface) is monitored at fixed time intervals.
RV I' performs the process of stopping the program only when the state is released, and if there is a D A SD that has not been released, cancels the non-stop m1. (Step 70
．． 71.72), IL E S E It V If the release of the E-shaped area is confirmed, the following 11'/! '＋M
Corresponding to 'Nrr, T Shita J-1/, S J month o (1
tj) (step 73), and when the memory saving is completed, the main 11VJ, machine 2 K R'
'j m knowledge J l (lj) 'A' Send! f
'Stop the operation of the W machine and turn on the power source lv1.

(Steps 74.75). .

In the case of ``J4i'' and ``J4i'', the name 9 [computer 4.4' from the child's notice 2.1f is constant! ai1. Is it received within the holiday period? Is it being monitored? Even if the monitoring time r has elapsed, a completion notification is returned.
Remove the fI machine from the system (After finishing monitoring the completion notices from all the subordinate g'iax machines of the Stesono 56L, the master n machine 2 initiates its own shutdown process yiJ+.

, that is, similar to the slave ifn machine, 几E S E IN
VI; After monitoring the status (steps 57, 58, 61) and saving the contents of the memory (step 59), stop the operation, turn off the power, etc.

In addition, unfreezing can be done by the main accountant n machine 2 and the wave calculation PA4.
4'0) Email contents 1JAsIJ8.1o, IO
'ff) i can be restored with the saved contents.

Effects of the Invention As is clear from the above explanation, according to the present invention, even in a complex computer system, a specific computer or all computers σ) can be forcibly stopped at any time, and moreover, it can be done centrally.
This has the effect of providing flexibility in system operation. In addition, it is possible to forcibly stop a computer system whose work load has decreased, such as during the night, without being affected by the progress of running jobs, resulting in automation of system operation, power saving, etc., and energy saving. It is effective in terms of

[Brief explanation of drawings]

Figures 1 to 3 explain one embodiment of the present invention. Figure 1 is a schematic configuration diagram of the system, Figure 2 is a flowchart showing the processing flow of the main computer, and Figure 3 is a flowchart of the processing flow of the slave computer. It is a flowchart which shows a processing flow. 1...Console device, 2...Payment J! Machine, 3...
- Stop processing task running on the main computer, 4.4...Slave computer, 5.5'...Channel coupling device (Ci'
CA), 6.6'...Stop processing task that operates on slave side computer, 7...DA shared between membrane number calculators.
SI), 810.10'... 1) ASD for saving memory contents, 11... Stop command from main computer to slave computer, 12... Completion notification from slave computer to master η1 machine. 'A-1 Figure 311 Figure 2

Claims (1)

[Claims] When a plurality of independent computers are connected to each other via an inter-channel coupling device, when a system M connection occurs, a specific computer (main computer) responsible for monitoring the entire system A stop command is sent to each other JV machine (slave computer), and each slave computer that receives this stop command starts a stop process that includes saving the contents of its main memory, and completes the process. Then, a completion notification is sent back to the main computer, and the accounting
One machine monitors whether completion notifications from each slave computer are returned within the monitoring time, and the slave computers that do not receive completion notifications within the monitoring time are separated from the system, and then the main computer stores its own main memory. A system freezing method characterized by starting a shutdown process including an operation to save the contents of the device.